EU and US to announce joint work on AI safety, standards and R&D

The EU and the United States are expected to announce cooperation in the field of artificial intelligence at a meeting of the EU-US Trade and Technology Council (TTC) on Friday, a senior European Commission official said while briefing reporters on background before the meeting.

The mood music demonstrates the growing collaboration between lawmakers on both sides of the Atlantic as they strategize to address the challenges and opportunities presented by powerful AI technologies — even as the business landscape remains deeply skewed, with U.S. giants like OpenAI continuing to dominate the development of cutting-edge AI .

The TTC was established several years ago after Trump’s election to provide a forum where EU and U.S. lawmakers could meet to discuss cooperation on transatlantic trade and technology policy issues. Friday’s meeting is the forum’s sixth since it began operations in 2021 and will be the last before elections in the two regions. The prospect of a second Trump term potentially undermining future EU-US cooperation is likely to focus lawmakers on maximizing opportunities for cooperation now.

“The TTC is definitely going to discuss the Office of Artificial Intelligence and [US] A senior European Commission official said the AI ​​Safety Institute was referring to an EU watchdog that is being set up as part of the upcoming EU Artificial Intelligence Act, a comprehensive risk-based framework using The framework for regulating artificial intelligence applications will start to apply across the EU later this year.

This element of the upcoming agreement appears to focus on AI safety or oversight, the official said, and is envisioned as a “cooperation or dialogue” between EU and U.S. AI regulators aimed at strengthening oversight of AI. Implementation of regulatory powers.

They said a second focus area of ​​the expected EU-US AI deal would revolve around standardization. This will take the form of joint work aimed at developing standards to support development by establishing an “artificial intelligence roadmap”.

The EU-US partnership will also have a third element, marked by “artificial intelligence for the public good”. According to the committee, this involves joint work to promote research activities, but with a focus on the implementation of AI technologies in developing countries and the global South.

The official said both sides had a common point of view: Artificial intelligence technology will be able to bring “very quantifiable” benefits to developing regions in areas such as healthcare, agriculture and energy. Therefore, this will also become a key area of ​​transatlantic cooperation to promote the application of artificial intelligence in the short term.

“AI” stands for Alignment of Interests?

As the EU says, artificial intelligence is no longer considered a trade issue by the United States. “Through the TTC, we are able to explain our policies and show Americans that, in fact, we share the same goals,” the commission official said. “Passing the Artificial Intelligence Act and [AI safety and security focused] Executive Order – Designed to reduce the risks of artificial intelligence technologies while supporting their use in our economy. “

Earlier this week, the United States and the United Kingdom signed a security cooperation agreement on artificial intelligence. Although the scope of EU-US cooperation appears to be broader – as it not only covers shared security and standardization goals, but also aims to coordinate efforts to promote the adoption of artificial intelligence in a range of third countries through joint support of “public interest” research.

European Commission officials revealed other areas of cooperation on emerging technologies, including standardization work in the area of ​​electronic identity (the EU has been working on e-ID proposals for several years), which they suggested would also be announced on Friday. They said that “electronic identity is a very strong area of ​​cooperation with great potential” and claimed that the United States is interested in the “huge new business opportunities” that will open up in the EU electronic identity wallet.

The official also said the EU and the United States were increasingly aligned on how to handle platform power, another area targeted by EU legislation in recent years. “We see a lot in common [between EU laws like the DMA, aka Digital Markets Act] Antitrust cases have also been launched in the United States recently,” the official said, adding: “I think there are undoubtedly win-win opportunities in many areas. “

At the same time, U.S. Commerce Secretary Gina Raimondo and British Technology Minister Michelle Donelan signed a U.S.-British Memorandum of Understanding on Artificial Intelligence in Washington on Monday. The two sides will work to accelerate joint efforts on a range of artificial intelligence security issues. work, including in the areas of national security and broader societal AI safety issues.

The U.K. Department of Science, Innovation and Technology (DSIT) said in a press release that the U.S.-UK agreement provides for at least one joint test of publicly accessible artificial intelligence models. It also suggested that personnel exchanges could take place between the two countries’ respective AI security agencies to cooperate in sharing expertise.

The U.S.-U.S. agreement envisions broader information sharing on the “capabilities and risks” associated with AI models and systems, as well as “underlying technical research on AI safety and security.” DSIT’s PR continued: “This will work to support a common approach to safe testing of AI, enabling researchers on both sides of the Atlantic and around the world to unite around a common scientific basis.”

Last summer, ahead of the Global Artificial Intelligence Summit, the British government said it had secured commitments from U.S. artificial intelligence giants Anthropic, DeepMind and OpenAI to provide “early or priority access” to their artificial intelligence models to support assessment and safety research. It also announced plans to set up a £100m artificial intelligence safety working group, which it said would focus on so-called foundational or cutting-edge AI models.

Meanwhile, at the UK AI Summit last November, on the heels of the US executive order on AI, Raimondo announced the establishment of a US AI Security Institute within a division of the National Institute of Standards and Technology, she said Aims to work closely with other AI security organizations set up by other governments.

So far, neither the US nor the UK has proposed comprehensive legislation on AI safety – the EU remains the leader in AI safety legislation. But more joint cross-border working seems a given.